[1.3.1] Ferram Aerospace Research: v0.15.9.1 "Liepmann" 4/2/18

[ferram4]

Matt77: Yes, I know.  It's been fixed in the dev build, continuously reporting the same issue doesn't make the dev release come out faster.  It actually makes it take longer because I have to check to make sure that this isn't a new issue.

SpacedInvader: That's an error in the part that has been safely (but not silently) recovered from.  Take it up with the part modeller, they're the one that added degenerate triangles to the mix.

Also, I can't reproduce that pFairings issue.  I can reproduce a different pFairings issue, where all the numbers end up as 0s and the log is spammed with update lines, but that is not related to that one and is fixed in the dev build.

[SpacedInvader]

11 minutes ago, ferram4 said:

SpacedInvader: That's an error in the part that has been safely (but not silently) recovered from.  Take it up with the part modeller, they're the one that added degenerate triangles to the mix.

Also, I can't reproduce that pFairings issue.  I can reproduce a different pFairings issue, where all the numbers end up as 0s and the log is spammed with update lines, but that is not related to that one and is fixed in the dev build.

Out of curiosity, what exactly is a degenerate triangle? 

As to the other issue, did you try removing the fairing and then adding it again? That seems to be where the issue crops up, as the correct curve / voxel placement is in place on the creation of the fairing, but then its removed and doesn't come back when reattaching the fairing again. On the other hand, it may already be fixed in the dev build compared to the release build, but I'm sure I was able to reproduce the issue on an install with only FAR and pFairings installed above stock...

More pictures in a step-by-step format:

http://imgur.com/a/svecD

[softweir]

7 hours ago, SpacedInvader said:

Out of curiosity, what exactly is a degenerate triangle? 

 

A degenerate triangle is one where:

• Two or more points are in the same place or
• The three points are in a line

In other words, a degenerate triangle has no surface area. If one of the points is a NaN (not a number) then that means the mesh was exported badly so one or more numbers are gibberish.

[Matt77]

22 hours ago, ferram4 said:

Matt77: Yes, I know.  It's been fixed in the dev build, continuously reporting the same issue doesn't make the dev release come out faster.  It actually makes it take longer because I have to check to make sure that this isn't a new issue

Thanks for the response.  I haven't continually reported the same issue and I've made no mention of release times.

Thanks for the excellent mod, though.

[sashan]

I've found that there's no need to have rockets aerodynamically stable as lon as you have high enough gimbal range, use SAS and keep angle of atack low while in dense atmosphere. That's teh way it's done in real life btw.

[blowfish]

16 minutes ago, sashan said:

I've found that there's no need to have rockets aerodynamically stable as lon as you have high enough gimbal range, use SAS and keep angle of atack low while in dense atmosphere. That's teh way it's done in real life btw.

That's generally been my experience too - particularly since FAR increases the gimbal range on all rocket engines.  I have needed fins on a few designs that are very draggy at the front, but generally full stability isn't necessary.

[ferram4]

Alright, so FAR v0.15.5.4 "Hoerner" is out, you should get it.  It includes a fix for the procedural fairings issues, what I suspect are almost all the causes of heatshielded pods having no drag during reentry, and a very rare crash that can occur during the initial aero loading.  I would say to check the changelog, but the forum is currently not letting me edit it, because it's so wonderfully functional.

[steve_v]

2 hours ago, ferram4 said:

...you should get it...

I'll trust you on that   I'd read back up the thread to see what's changed, but holy mackerel it's unpleasant to look at now.

Edited December 2, 2015 by steve_v
Effing horrible editor. Grrrrr!

[steve_v]

I'm having issues with control deflection on the delta-deluxe winglets, seems like it's only a fraction of what I set in the tweakable.

All other control surfaces I've tested appear to be OK, but I can't get more than a few degrees out of the delta deluxe...

40 degrees deflection that is not, and it's the same in the flight scene. Ditto when set for spoilers / flaps.

Clean 1.0.5 + FAR 0.15.5.4 install.
Full logs available on request of course.

Recon you could take a quick look?

 

Edited December 2, 2015 by steve_v

[kcs123]

55 minutes ago, steve_v said:

I'm having issues with control deflection on the delta-deluxe winglets, seems like it's only a fraction of what I set in the tweakable.

All other control surfaces I've tested appear to be OK, but I can't get more than a few degrees out of the delta deluxe...

40 degrees deflection that is not, and it's the same in the flight scene. Ditto when set for spoilers / flaps.

Clean 1.0.5 + FAR 0.15.5.4 install.
Full logs available on request of course.

Recon you could take a quick look?

 

I was been confused about it in earlier FAR release until ferram explained it.
When you set desired angle and run AoA graph analyse, it shows visualy on control surfaces deflection angle on latest(rightmost) AoA.

Meaning, visual representation of control surface angle is calculated based on rightmost AoA - especialy important info if you also have AoA% set on some surfaces.

I'm not sure what was the case when you don't have AoA% and you set pitch settings on "1", like on your picture, though.
Regardles of AoA%, you should run analyses in either case to see visual representation of control surfaces. It is not clear from picture if you have run graph analyses or not when you have changed control surface deflection.

I hope this info helps.

[steve_v]

33 minutes ago, kcs123 said:

It is not clear from picture if you have run graph analyses or not when you have changed control surface deflection.

How 'bout this one, both control surfaces are set to 40 degrees deflection:

 

[BigNose]

Nice update, thanks ferram4!

[ferram4]

Steve_v, that's just a visual thing on parts that don't have the control surface making up the entire surface.  It's a simplification to reduce runtime calculations (so that I don't have to recalc a new angle just for the deflection displayed).  Physics are correct, visuals are just weird, as with all legacy things.  Wing overhaul should fix that when it comes around.

[steve_v]

Excellent, a little confusing but no big deal. Thanks for clearing it up, I'll stop worrying and go back to enjoying another top quality FAR release.

[sashan]

What we all really need is a way to make proper working Fowler flaps without IR and a dozen parts for each section.

 

[Kagame]

So what can we expect with the wing update? Would I finally be able to implement fences, wingtip devices, and slotted flaps?

[ratchet freak]

Would it be possible to have a set of sliders (airspeed, pitch, bank, yaw, air density, etc.) to move the CoL indicator?

This would allow a "simpler" visualization of how the plane would behave for those who can't figure out how to read the graph.

For example if you set up the cruise configuration of a plane and you increase the speed then you would immediately see how that increased speed affects the aircraft.

Then as an extra if you do an AoA analysis and you click on the graph the relevant parameters are selected and the CoL ball moves to show where the lift force ends up and how strong it is.

[tetryds]

@ratchet freak but that is exactly what the graphs do, you would be only limiting yourself to reading a single point at a specific environment, it means less than what you may think it does.

If someone cannot read the graphs I don't know how implementing something complex like that would help, everytime someone posts about this COL thing I get one step closer to bringing up that it should be removed.

And besides, you already have what you are asking, it's the Mw on the derivatives table, the more negative it is, the further from the center of mass the vertical force is acting, it's just a value rather than a point.

 

Edit: banking your airplane on the SPH works, by the way.

Edited December 3, 2015 by tetryds

[steve_v]

The CoP (it's not CoL BTW, there's way more going on than just lift) marker isn't all that meaningful anyway - IMO it's still handy in the early stages of a layout, but it's so not-indicative-of-actual-performance that Ferram has threatened to remove it altogether at least once before.

The real issue here is that the simulation is too complicated (read realistic ) to represent "lift" as a point (let alone a vector, as in stock) - so while it's probably possible to do what you describe, it would have very little bearing on flight characteristics.

In short: learn how to use the analysis tools provided, they already do what you're asking for, they're way better than the simplified CoP ball, and your aircraft designs will thank you for it.

E.g. Need to know what AoA is required for level flight at a particular altitude and speed? Want to know how that AoA influences drag, and whether you have enough thrust to maintain it without loosing velocity? The analysis tools will tell you - the CoP ball most certainly won't.

Edited December 3, 2015 by steve_v

[kcs123]

@ratchet freak
I haven't updated my thread for a while, info there is based on older version of FAR, but data provided there is still valid.
I have prepared some new pictures covering area ruling and other new FAR features, but real life issues kicked in and haven't found enough time or willpower for proper edit that thread.

Check out links in my signature and picture galeries if you have trouble how to use FAR graph and analysis.

[The DigitalAlchemist]

Hey, I'm having an issue with all of my aircraft that after a certain speed they start osculating wildly and break apart. I cant maintain stable flight for any period of time. If I disable aerodynamic failure, the oscillation just gets worse until I lose control of the aircraft. 

 

In stock, this was a very good, very stable aircraft even at speed. Very easy to control. 

 

I've had this issue in previous versions too, it's the single thing keeping me away from using it. I like the better physics but having my craft wobble about for no apparent reason is very obnoxious.

 

Further, I'd like to know if it's possible to make aerodynamic failures happen close to KSP default values WITHOUT turning it off all together. I dont think my planes should be falling into shrapnel at 150 m/s because I tried to tweak my pitch. Default, any maneuver at mach speed is suicide (I blame it on digital controls). 

[Wanderfound]

2 hours ago, The DigitalAlchemist said:

Hey, I'm having an issue with all of my aircraft that after a certain speed they start osculating wildly and break apart. I cant maintain stable flight for any period of time. If I disable aerodynamic failure, the oscillation just gets worse until I lose control of the aircraft. 

 

In stock, this was a very good, very stable aircraft even at speed. Very easy to control. 

 

I've had this issue in previous versions too, it's the single thing keeping me away from using it. I like the better physics but having my craft wobble about for no apparent reason is very obnoxious.

 

Further, I'd like to know if it's possible to make aerodynamic failures happen close to KSP default values WITHOUT turning it off all together. I dont think my planes should be falling into shrapnel at 150 m/s because I tried to tweak my pitch. Default, any maneuver at mach speed is suicide (I blame it on digital controls). 

 

These are more questions for the design thread: http://forum.kerbalspaceprogram.com/threads/121176-Official-FAR-Craft-Repository

If you post some screenshots and analyses (see http://forum.kerbalspaceprogram.com/index.php?/topic/81848-kerbodyne-ssto-division-omnibus-thread/&do=findComment&comment=1341606 if you don't know how to do these) over there, we'll be able to help you fix your design.

Wing strength is now tweakable part-by-part, and the default strength is more than enough for most circumstances. Hair-trigger wing breakage hasn't been a standard FAR thing for a long time. If you're still having those problems, there's likely to be something seriously wrong with your airframe or control surface setup.

[SpacedInvader]

On 12/3/2015, 4:39:05, steve_v said:

The CoP (it's not CoL BTW, there's way more going on than just lift) marker isn't all that meaningful anyway - IMO it's still handy in the early stages of a layout, but it's so not-indicative-of-actual-performance that Ferram has threatened to remove it altogether at least once before.

The real issue here is that the simulation is too complicated (read realistic ) to represent "lift" as a point (let alone a vector, as in stock) - so while it's probably possible to do what you describe, it would have very little bearing on flight characteristics.

In short: learn how to use the analysis tools provided, they already do what you're asking for, they're way better than the simplified CoP ball, and your aircraft designs will thank you for it.

E.g. Need to know what AoA is required for level flight at a particular altitude and speed? Want to know how that AoA influences drag, and whether you have enough thrust to maintain it without loosing velocity? The analysis tools will tell you - the CoP ball most certainly won't.

Is there a guide somewhere describing how to read the analysis tools in relation to rockets? I know how to use them for aircraft / spaceplanes, but have little to know understanding of how those principles apply to long skinny tubes meant to get out of the atmosphere as quickly as possible...

[ferram4]

20 hours ago, The DigitalAlchemist said:

 

Hey, I'm having an issue with all of my aircraft that after a certain speed they start osculating wildly and break apart. I cant maintain stable flight for any period of time. If I disable aerodynamic failure, the oscillation just gets worse until I lose control of the aircraft.

That sounds realistic actually; your design is highly flexible and extremely prone to flutter.  I would suggest many more struts as a solution.  Real planes flex a lot unless you take steps to fix it.

 

Alternatively, decide that the stock game's joints are terrible and go for Kerbal Joint Reinforcement.  From the way things have seemed to me, they've backed off the joint stiffness a bit every update or so, so things get a lot more wobbly again.

 

20 hours ago, The DigitalAlchemist said:

Further, I'd like to know if it's possible to make aerodynamic failures happen close to KSP default values WITHOUT turning it off all together. I dont think my planes should be falling into shrapnel at 150 m/s because I tried to tweak my pitch. Default, any maneuver at mach speed is suicide (I blame it on digital controls). 

Well, it depends on how much load you're putting on that control surface.  You're doing a real impressive job if you're getting the default wing strength parts to come off at 150 m/s at SL; that implies you're getting a lift coefficient of 2.8... okay either you're reducing the wing strength parameters too much or your install is borked.  By default, each wing should be able to take ~40 kN/m^2 of lift force before it fails.

 

8 hours ago, SpacedInvader said:

Is there a guide somewhere describing how to read the analysis tools in relation to rockets? I know how to use them for aircraft / spaceplanes, but have little to know understanding of how those principles apply to long skinny tubes meant to get out of the atmosphere as quickly as possible...

For AoA sweep, you're really only interested in the yellow line (Cm).  As usual, a negative slope means stable, but you're unlikely to get that.  Instead, just make sure that it's not that stupidly large a positive slope.  Remember to check at different Mach numbers, particularly around 1 - 1.6, since that's where Max Q is likely to be for you.  For a Mach sweep, just make sure that the red line (Cd) doesn't spike way too much near Mach 1.  If it does, you might want to smooth out the vehicle a bit.

Remember to do this for all the configurations that will have strong aerodynamic forces on them, so with / without boosters, and possibly the second stage if the first stage drops off very early.

[kcs123]

For rockets in FAR, another good tip is to reduce engine max. thrust in hangar, so it does not accelerate rocket too quickly.
Try to tweak engines in a way that you breach mach 1 only when you are near 10km of altitude. It also helps that you don't do gravity turns until you reach less dense atmosphere (again after 10km).

Using engines with good gimbals for stages up to 30 km also helps and when you start to steer rocket make sure that AoA is low enough - you need to wait for a while until prograde vector align with direction where you want to go. Otherwise you may turn rocket in unrecoverable spin due to too large AoA.

Like Wanderfound said, for more questions about designs better ask in FAR exchange thread, to help making this thread more "clean" so bugs reports and similar issues does not left unnoticed.